OpenGeofiction

Luciano's Mapper's Challenge #5 - August, 2016 - WHAT'S IN A NAME?

Posted by Luciano on 2 August 2016 in Korean (한국어)

For my challenge last month, "Let's do something blue," I think I did pretty well. Mostly, I'm proud of the Irhoborin Refugee Settlement, in Duvar, Commonia. I also did some work on administrative divisions in various blue countries, and I laid down a first draft for the Kshang Native Lands Area in Rhododactylia - but that last still needs a LOT of work.

How did you do? I saw some people doing some great stuff in blue areas this past month.

Actually, the main thing I worked on this past month has been my little city-state of Tárrases. I will post about that separately, at some point.

Thinking about names

I wasn't sure what to make as a challenge this month. So I decided to just think about what I would be working on.

For me, names on a map are important. Nothing is more disappointing for me than to see an area that looks well-mapped, realistic, and interesting, only to zoom in and find that most of the streets and other objects have no names attached to them. It detracts from the realism of the map, and makes the whole thing feel more like a kind abstract art that happens to be map-like. I know names are not interesting to everyone - if names are not your thing, then this challenge is not for you.

I try very hard to name things on my maps as I build them, but I'm certainly not perfect in this respect. Sometimes my names can become repetitive or unimaginative - but sometimes names in the real-world are repetitive and unimaginative, too. How many places bear the name "Washington" in the US? 100s if not 1000s. How many places bear the name "Santa María" in Latin America? How many places bear the name "Gwangju" in Korea?

Anyway, the challenge this month is to try to increase the number of names on your map, and improve the quality, consistency and "story" behind your naming schemes, whether based on real-world languages or your own invented language.

"Name Density" - toward an objective measure of map quality

I have been trying to develop a concept of "name density" - the number of named objects per square kilometer. Here is an example.

I downloaded a typical, well-mapped area of OSM, in the northeast corner of the Basque Country in Spain. It's a small area, but it's a rectangle that's roughly 16.31 sq km, including about 10% water, with both urban and rural features.

I used osmfilter's statistics function, and very interestingly I got 1631 names. That means exactly 100 names/sq km.

That was such a suspcious first result that I grabbed an .osm file I have of Mexico City, Delegación Cuauhtemoc, the neighborhood where I used to live. It's a bit bigger, being 44.56 sq km, all urban. It has few open areas at all compared to the first area I did, but the quality of the mapping is also much less thorough.

I used the statistics function on this file and got 2793 names per sq km, or a name density of about 63 names/sq km.

After that, I got carried away, and I spent several hours doing a bunch of different areas. I also looked at some OGF locations. Here is a screenshot of the spreadsheet I made.

spreadsheet

Some observations:

Obviously, the quality of mapping might also influence the name density of these areas. My current home in Ilsan, Korea, is a good example of this. There are probably a million residents in the enclosed area that I downloaded, yet the OSM mapping quality is notably sparse.

More importantly, however, is to remember that an urban area will obviously have a higher name density than a rural area. It might be more interesting to come up with a ratio of population density to name density. This would actually be a kind of indicator of map completeness and map quality. But it's beyond my current ability to do this, because it means getting downloaded OSM areas that match known population figures with some precision. I think it's an interesting goal.

For now, let's keep things simpler than that.

Luciano's Mapper's Challenge #5 - August, 2016 - WHAT'S IN A NAME?

As you can see, the last area I did in my spreadsheet was my current obsession, the city state of Tárrases. The result that I got, at 29 names/km sq, is about in line with what I would expect for this stage in my work there. I was not disappointed at all. But it gave me an idea for a goal for my challenge for this month.

The total land area of Tárrases is 203 km sq. My goal for this challenge is to raise the city-state's name density to 100 names/sq km. That means there need to be about 20,000 named objects on the map. I'm going to have to put in a lot of little shops and restaurants!

If you want to attempt this challenge, choose some area (whatever size you're comfortable working with), and try for a "real world" level of name density - whatever is appropriate for the type of mapping style and type of area (rural or urban) you're working on.

Happy mapping, and happy August

Comment from Yuanls on 2 August 2016 at 08:13

This would sound fun if I hadn't already mapped in nearly the highest detail possible like I've done for Harcourt.

Hide this comment

Comment from Luciano on 2 August 2016 at 08:45

@Yuanls - I downloaded Harcourt and ran the numbers. I respected the area with landuse/landcover definitions (ignoring "gray land" since it's clearly not done yet).

215 names / 14.45 km sq = 14.88 names / km sq

That's not bad, especially for what is clearly a rural area. Pretty much on target; anyway, better quality data than e.g. Ketchikan.

Hide this comment

Comment from Demuth on 2 August 2016 at 09:06

This is perfect timing for my current project: I'm trying to create a fairly detailed urban/rural mix area around the large town of Aneborn in south-central Østermark. I'm putting in buildings in the town at the moment, and while I've named some streets, I hadn't thought so much about how much giving names to streets adds to the quality of the map. So that's going to be my focus for this challenge, naming all the streets (and also other details of course) as I work.

Hide this comment

Comment from Ūdilugbulgidħū on 2 August 2016 at 09:53

I've really no idea how I did on the last challenge. I was sort of all over the place as I tried to tidy up past mistakes and things I'd partially mapped before. The Hsru lands got a bit more mapped, but still only one village there. A few buildings appeared in the blue countries. I don't think I made anything worse than it was before.

This one is a very interesting challenge - Luciano, can I ask you to 'run the numbers' on a couple of places so I can decide what to focus on?

Also, does anyone know how exactly the name search function works in OGF? Places can be found, but it seems only possible if the place is within the viewing window. This is different from OSM - makes it much harder to find even 'well known' places.

Hide this comment

Comment from Demuth on 2 August 2016 at 09:56

p.s. You've lived in some very different and far-flung places, Luciano. What a combination! Makes me very curious...

Hide this comment

Comment from BMSOUZA on 2 August 2016 at 10:08

Another very interesting challenge! This one I can do, hehehe!

As I am also with an "obsession" these days, of course I will do the challenge in Slovech.

Currently I am focused about creating relations to bus and tram routes in Slovech area. After this, I will continue filling the city with buildings, and names in the streets.

Naming Bolgrad first could be easy - Luciano, how do you count the names in an area?? (Sorry if you wrote about it, I dont understand english perfectly yet, hehehe).

Hide this comment

Comment from BMSOUZA on 2 August 2016 at 10:09

PS: As Demuth, I am curious too, hehehe

Hide this comment

Comment from Luciano on 2 August 2016 at 10:23

I will be making an OGF: Help page in the next few days showing how you can use JOSM and the osmfilter utility to calculate your own name density. Once I've done that, I'll also take requests for people who don't want to take that on. Or if you message me a request, I'll get around to it when I can - I can't always do it quickly though.

Hide this comment

Comment from wangi on 2 August 2016 at 13:52

If folk want to take a stab at doing their own stats before Luciano can write it up, then here is a rough guide:

  1. In JOSM draw a polygon around your area / city; using the measurement plugin you can see the area: http://wiki.openstreetmap.org/wiki/JOSM/Plugins/measurement

  2. Save the polygon as a .poly file

  3. Use command line osmconvert tool with the -B option to clip your .osm file to the city polygon .poly file: http://wiki.openstreetmap.org/wiki/Osmconvert#Clipping_based_on_a_Polygon

  4. Use the command line osmfilter tool with the --out-count option to calculate the number of "name" tags in the clipped file: http://wiki.openstreetmap.org/wiki/Osmfilter#List_of_all_Keys.2C_sorted_by_Occurrence

Hide this comment

Comment from Luciano on 2 August 2016 at 14:05

@wangi - spectacular. Much clearer and more efficient than I could have phrased it. As one can see, I've never been very good at brevity.

Hide this comment

Comment from No Way on 2 August 2016 at 14:32

Can the calculation be done in sq mi in lieu of km sq?

Hide this comment

Comment from wangi on 2 August 2016 at 15:56

Knock yourself out: 100 ha = 1 km2 = 0.3861 mile2

Hide this comment

Comment from _zM on 2 August 2016 at 20:11

I think I did okay in last month's challenge. I only did very little, but that was because of multiple different reasons (school etc.) My work consisted mostly of filling in empty spaces in Gobras City with residential areas and adding exits to the Motorway 3 in northern Yantia District. Then there was some work done on the subways in Gobras City.

For this month I guess I'll do the core city of Akesia. The main infrastructure is in place, now there is only some detailing to do. I might also improve Noian-Piruna a bit.

Hide this comment

Leave a comment

Parsed with Markdown

  • Headings

    # Heading
    ## Subheading

  • Unordered list

    * First item
    * Second item

  • Ordered list

    1. First item
    2. Second item

  • Link

    [Text](URL)
  • Image

    ![Alt text](URL)

Login to leave a comment